This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. Autism represents one of the most severe disorders of development, yet its biological underpinnings remain unknown. Although the prevalence of autism in the general population is estimated at 0.05 -- 0.1 percent, a variety of studies indicate that deficits in social functioning and communication affect 15 percent of first-degree relatives of autistic individuals. The poor long-term outcomes seen in such children underscores the need for a better understanding of the biological mechanisms involved in autism. Three intersecting domains of behavior: communication, social reciprocity and restricted/repetitive behaviors, taken together are diagnostic of the illness, but appear partially independent. This fact, in concert with evidence from genome-wide linkage investigations and candidate gene studies, provides evidence that a single etiology of autism is unlikely. Given the heterogeneous nature of autism, identification of homogeneous subgroups is essential to furthering our understanding o f the biology underlying the behaviors associated with this disorder. The availability of noninvasive, non-radioactive neuroimaging techniques and sophisticated data analytic approaches used in combination with family genetic data holds the promise of greatly improving our ability to identify these subgroups. The principal goal of the work proposed in this project is to use these tools to identify neuromorphometric and neurochemical measures that define meaningful endophenotypes in autism. In addition, the extent to which these measures are shared among siblings and therefore are familial will be explored as a means of identifying these endophenotypes. Using these tools, Affected Sibling Pairs selected from the Autism Genetic Resource Exchange multiplex families, and a sample of non-autistic control children selected from an ongoing study of normal children at UCLA will be examined. The CAN/AGRE study population is a large group of multiplex families with phenotypic and genotype data a vailable for access by research scientists. We propose that key behavioral phenotypic dimensions of autism including: delay of language onset, repetitive/stereotyped behaviors and social reciprocity, are associated with specific brain neuromorphometric and neurochemical measures. Furthermore, we will examine the extent to which the neurobiologic findings underlying these phenotypic variations will be correlated among siblings, suggesting that they may be useful phenotypes for future investigations. Taken together, these studies will provide the basis for genetic studies to delineate specific gene brain behavioral pathways and lay the foundation for early detection and better intervention in autism.